Many RF signal processing applications require that an amplitude modulated RF signal be applied to a load, such as, but not limited to an acousto-optic modulator, whose known power handling capabilities may be exceeded under conditions where the total energy delivered in a given period of time may cause damage to the device. For the most part, damage may be attributable to excessive heat generation, which must not be permitted to exceed a maximum safe level in order to ensure proper operation of the device. At the same time, the instantaneous power level delivered to the load may be quite large, as in the case of intermittent pulse radar or low duty cycle drive requirements for certain acousto-optic modulator (AOM) applications.
In a typical AOM application, a reduced complexity diagram of which is illustrated in FIG. 1, a source of RF carrier power is coupled by way of an RF input port 11 to an input gain stage 12. The output of the input RF gain stage is coupled by way of a variable gain attenuator or amplifier (VGA) 13 to a first, RF carrier input port 14 of an amplitude modulator stage 15. Amplitude modulator stage 15 has a second, baseband input port 16 to which a baseband modulation signal Vam is supplied. The output of the baseband modulator 15 produces a modulated RF signal which is amplified in a main RF power amplifier 17 and then applied to an output device or load, such as an acousto-optic modulator 18. If the baseband signal Vam is not controlled by the user in such a manner as to limit the amount of RF energy delivered to the load within a given time interval (i.e., in a manner effectively limiting the average RF power level over that interval), damage to the load may occur.
Conventional approaches for limiting RF power rely on direct measurement of the instantaneous modulated RF envelope being supplied to the load, with post integration of the detected RF power level. Post integration is most often performed by means of a low pass filter network, whose output depends on the impedance of the RF power detector and the baseband spectrum, both of which contribute to reduced output with very short duration pulses (e.g., less than ten nanoseconds). As these approaches are expensive, there is a need for a more ‘practical’ (i.e. cost effective) mechanism for limiting RF output power to safe levels, that will not induce thermal overload in the output device.
Theory of Operation
Referring to FIG. 1, let the source RF carrier power level of RFin be given by Pin. Further let GA be the product of all constant RF gain sources in the RF signal path, with variable gain GV from a variable gain attenuator or amplifier, and a gain GM associated with the RF modulator. Then the total output energy Eo supplied to the load in a time τ is given by:                               E          o                =                              ∫            0            τ                    ⁢                                                    P                o                            ⁡                              (                t                )                                      ⁢                                                   ⁢                          ⅆ              t                                                          (                  E          .                                           ⁢          1                )            where Po(t), the instantaneous output power delivered to the load, is just the product of Pin and the total gain. Rewriting (E.1) in terms of the product gains gives:                               E          o                =                                            G              A                        ·                          G              V                        ·                          Pi              n                                ⁢                                    ∫              0              τ                        ⁢                                                            G                  M                                ⁡                                  (                  t                  )                                            ⁢                                                           ⁢                              ⅆ                t                                                                        (                  E          .                                           ⁢          2                )            
Further, if GM(t) and VAM(t) are related to within a suitable scaling constant and offset, ie. a linear relationship exists between the two, then ignoring the offset:GM(t)=κ·VAM(t)  (E.3)and,                               κ          ·                                    ∫              0              τ                        ⁢                                                            V                  AM                                ⁡                                  (                  t                  )                                            ⁢                                                           ⁢                              ⅆ                t                                                    =                              ∫            0            τ                    ⁢                                                    G                M                            ⁡                              (                t                )                                      ⁢                                                   ⁢                          ⅆ              t                                                          (                  E          .                                           ⁢          5                )            
The leftmost integral may be determined directly in any circuit application by use of a simple analog video integrator, where the output voltage from the integrator Vo is given by:                               V          o                =                                            -                                                A                  V                                                  τ                  o                                                      ·                                          ∫                0                τ                            ⁢                                                                    V                    AM                                    ⁡                                      (                    t                    )                                                  ⁢                                                                   ⁢                                  ⅆ                  t                                                              =                      C1            ·                                          ∫                0                τ                            ⁢                                                V                  AM                                ⁢                                                                   ⁢                                  ⅆ                  t                                                                                        (                  E          .                                           ⁢          5                )            with a suitable video scaling constant Av applied, where τ0=R·C is the integrator time constant, given by the product of the input resistance R and feedback capacitance C. Note there are no restrictions placed on the form of VAM(t), other than its suitability for application to a video integrator. From (E.4) and (E.5) we obtain the result:                                           κ            C1                    ·                      V            o                          =                              ∫            0            τ                    ⁢                                                    G                M                            ⁡                              (                t                )                                      ⁢                                                   ⁢                          ⅆ              t                                                          (                  E          .                                           ⁢          6                )            
Correspondingly, if GM(t)=Gmax is constant over the integration period τ, where Gmax represents the maximum gain associated with the modulator, then:                                           κ            C1                    ·                      V            max                          =                                            ∫              0              τ                        ⁢                                          G                max                            ·                                                           ⁢                              ⅆ                t                                              =                                    G              max                        ·            τ                                              (                  E          .                                           ⁢          7                )            where Vmax represents the maximum extrapolated output voltage which the integrator would produce. Dividing both sides of (E.2) by the integration period τ, gives the result:                               P          avc                =                                                            G                A                            ·                              G                V                            ·                              Pi                n                            ·                              1                τ                                      ⁢                                          ∫                0                τ                            ⁢                                                                    G                    M                                    ⁡                                      (                    t                    )                                                  ⁢                                                                   ⁢                                  ⅆ                  t                                                              =                                    G              A                        ·                          G              V                        ·                          Pi              n                        ·                          〈                                                G                  M                                ⁡                                  (                  t                  )                                            〉                                                          (                  E          .                                           ⁢          8                )            where Pave is the average output power delivered to the load during the integration period, and <GM(t)> is the time average gain of the modulator over the same time interval. Using (E.6) and (E.7), we can express this result in terms of the integrator output voltages as:                               P          ave                =                              G            A                    ·                      G            V                    ·                      G            max                    ·                      P            in                    ·                      (                                          V                o                                            V                max                                      )                                              (                  E          .                                           ⁢          9                )            
This is the key result as applied to the current claim, wherein a simple circuit concept is proposed which, utilizes the established gain values for GA, GV, Gmax and Vmax, measures Pin and performs the necessary integration of the modulation signal in order to set Pave to a safe operating level.
By way of example, we take the case of simple pulse modulation, where we have the following conditions:                                                                                           G                  M                                ⁡                                  (                  t                  )                                            ;                                                                    V                    AM                                    ⁡                                      (                    t                    )                                                  =                                ⁢                                                                            G                                              M                        ;                                                              ⁢                                          V                      '                                        ⁢                                                                                   ⁢                    for                    ⁢                                                                                   ⁢                                          V                      i                                                        ≥                                      V                    th                                                                                                                                        ⁢                              0                ;                                                      0                    ⁢                                                                                                               ⁢                                                                                                             ⁢                    for                    ⁢                                                                                   ⁢                                          V                      i                                                        ≺                                      V                    th                                                                                                          (                  E          .                                           ⁢          10                )            where the value of the modulating voltage is V′, for an input control voltage Vi which exceeds a logic threshold, and zero otherwise. Then (E.8) gives;                               P          o                =                              G            A                    ·                      G            V                    ·                      G            M                    ·                      P            in                    ·                      (                                          τ                on                            τ                        )                                              (                  E          .                                           ⁢          11                )            where τon is the total “on” time of the modulator during the integration interval τ, during which it exhibits a maximum constant gain GM. By performing the corresponding video integration of the modulating voltage as expressed by (E.6) and (E.7), we obtain the result;                                           τ            on                    τ                =                              V            o                                V            max                                              (                  E          .                                           ⁢          12                )            which upon substitution in (E.11) produces the same results given by (E. 9) for the general case. The approach to providing a low cost energy limiting function which fully exploits the results of the theory described, forms the basis for the present invention.